Music files for big stations?

[Groove1670]

Curious. What type music files do the big C companies use. .wav or mp3. Sometimes that audio quality is hit or miss depending on the station.

 

[ChiefOperator]

Most of the stations I'm contracted for are moving to PCM. Just a few years ago, the most popular seemed to be MP2@256 or 384. I know that the CC cluster nearby is using PCM and MP2. I don't know of any stations that have standardized on MP3.

 

[Kent T]

PCM is fast becoming standard. MP3 is not suitable for broadcast. Especially with modern digital airchains. Cascading algorithms tend to lead to artifacts. Which engineering won't approve of. A lot of modern recordings are very dynamic compressed and mastered at too hot levels. And unfortunately sound like garbage.

 

[TomT]

.wav No sense using compressed formats with the large hard drives. Either loaded off CD or from TM download (for the older cuts).

Automation will play either MP2 or MP3, but we use these formats mostly for spots and speech cuts. Of course,most spots are downloads anymore. Weather forecasts are shipped back and forth by MP3, some locally down, some from the service.

 

[Groove1670]

MP3 is not suitable for broadcast.

I understand that. But I am willing to bet that there is one large station or group (in a selected market) out there somewhere with several songs on the playlist with MP3 files.

 

[frankberry]

If the station has analog audio processing and analog STL, high bit rate MP3 files would probably be fine.
The problem occurs with digital processing and digital STLs because they use some form of compression for bit rate reduction too.

 

[ChiefOperator]

MP3 is not suitable for broadcast.

I understand that. But I am willing to bet that there is one large station or group (in a selected market) out there somewhere with several songs on the playlist with MP3 files.

No doubt about that. I engineer a station that uses iMediatouch. This program uses a transcoder that encodes any import into PCM. There is a production guy who exports out as mp3 and then imports again as PCM. He may do this several times on the same file so that the file that airs has been encoded 5 or 6 times. I've explained the problem to him, yet his comment is that I'm out of touch because "mp3 is the standard." You can hear his files on the air and when they stream, they are a total train wreck.

Ad agencies send their files in compressed format, usually mp3@128. However, some send their files as mp3@320 or mp2@384. When I see these higher rates, then I know that the production guy knows his stuff.

Have a look at this:

http://boards.radio-info.com/smf/index.php?topic=198783.0

 

[Groove1670]

Ad agencies send their files in compressed format, usually mp3@128.

Several spots that air on syndicated talk shows sound awful, probably sent out @ 128.

Have a look at this:
http://boards.radio-info.com/smf/index.php?topic=198783.0

That was a great thread. Spot quality is a hit or miss (you sometimes can control what is sent to you). But music........

Why spend a bunch a money on processing if......

 

[Kmagrill]

Bit rate reduction, often called "compression" when referring to lossy codecs, is different from audio compression. Audio compression is simply the process of dynamically changing the gain to make all sound come out at about the same amplitude. A digital audio processor does this without throwing data bits away, so there's really no problem of dualing algorithms, per se. High quality MP2 and MP3 files both seem to work fine on a digital processor, as near as I can tell. If they didn't, most of your satellite delivered programming, which is bit-rate-reduced, would have problems on the air. Low quality mp2/mp3 files should be avoided, however since compressing the audio is like putting it under a magnifying glass. Any defects are more likely to be heard, so it's important that the audio not have any tiny audible defects to amplify. When using mp2/mp3 files, what often gets stations into trouble is when some form of lossy compression (bit rate reduction) is applied to the path somewhere between the automation output and the transmitter, like when a compressed (bit rate reduced) STL is in the path. Then the dualing algorithms can play havoc with the sound. Likewise, changing sample rates can have an audible effect, similar to bit rate reduction and seems to affect mp2 & mp3 files more than PCM audio. Another important factor is if the station streams. Since streams are compressed, there will be dualing algoithms if mp2 or mp3 files are used as streaming sources. And don't forget that HD radio is just another form of AAC streaming.

Perhaps one of the audio processing gurus can chime in with better information than I. I'd be interested to hear what they think on this subject.

In any case, PCM audio is preferred and if bit rate reduced music is unavoidable, it should be at the highest possible data rate. For mp2, nothing lower than 256 must ever be used (384 preferred) and either 256 or 320k are preferred for mp3. Never assume that converting an mp2 or mp3 to PCM will avoid problems. Once the audio has been converted to a lossy format, the damage is irreversible. You cannot put the genie back in the bottle and converting from one lossy format to another is the audio kiss of death.

 

[Groove1670]

I will keep my 8100 & XT2 as my audio chain. Analog. Simple.

 

[chriscollins]

We use all PCM .wav, except for the emailed commercials (mp3).

Your audio processor does not matter. They operate linear and will not bit reduce your content. Now, certain digital STL's WILL bit reduce. The MPEG2 streams coming down satellites in regards to morning shows with music, etc... are at a very high bit rate. They can be bit reduced down to stream level with no additional artifacts.

 

[Goat Rodeo Cowboy]

Your audio processor does not matter. They operate linear and will not bit reduce your content.

I can't say that you are wrong, for I really don't have hands-on experience with one, but from reading what "the big-boys" have to say over in the technical forums and in some other locations, SOME audio processors may be digital. The fact that consoles are now digital can also mean that adio may be at one bit rate in HD storage, converted to another bit-rate in the mixing console, convertered to another bit-rate and/or big depth in processing at the studio, another digital domain in the STL and yet another digital protocol in the final peak limiting at the TX site, and to further pile-on, the transmitter exciter may be of yet another "digital religion".

Some of the debate seems to be ag'd on by purist who would like to see NO digital involved, and counter-punches are thrown by people who assure us that digital processing and conveyance can NEVER be a problem.

Your mileage may vary.

 

[chriscollins]

There can be compression in Digital STL's... Typically when someone is cramming multiple stations to a common site. We use some uncompressed digital links.

I promise you that Optimod & Omnia do absolutely NO compression.

In a facility, it is doubtful that there are multiple sample rates. Everything is going to be clocked to a standard reference clock (for me it's 44.1). Automation systems may perform this duty for you, but it would be an upconversion on very old files only. Back in the day 32Khz, MPEG2 was common when hard drives were expensive.

I work for a six station cluster. If we were able to go back and get everything at 44.1 Uncompressed, I am sure Cumulus and CC either have or are in the process of doing so. Streaming is VERY important to them and they are well aware of the crap you get when a low bit rate lossy song passes through an even lower bitrate AAC codec.

My whole chain is digital, BTW.

 

[davideduardo]

I don't know of any stations that have standardized on MP3.

Funny, because there is a de facto use of MP3s since many record companies distribute new releases first as MP3's. Some, later, make .wav files available later, but not everyone gets them and many stations just keep playing the original MP3 release. ..

 

[xmusicmatt]

The MPEG2 streams coming down satellites in regards to morning shows with music, etc... are at a very high bit rate. They can be bit reduced down to stream level with no additional artifacts.

The stuff on satellite from Dial Global etc.. is all still compressed.. If I am not mistaken they use 192kbps MPEG2... at least I think they did in the star guide days.

 

[ChiefOperator]

I don't know of any stations that have standardized on MP3.

Funny, because there is a de facto use of MP3s since many record companies distribute new releases first as MP3's. Some, later, make .wav files available later, but not everyone gets them and many stations just keep playing the original MP3 release. ..

As I said, I don't know of any station that has standardized on MP3. Which music service distributes in only mp3 and which station has standardized on that format as a result? As I said, I don't know of one, although I obviously don't know the formats of all stations. I've certainly seen music services use MP2, as I've already stated. In fact, MP2@256 was used by CC for many years. Don't know if it still is.

 

[xmusicmatt]

Many years ago (back in 1999 when I worked for Clear Channel.. We used MP2 256kbps files on the "Audio Wizard" system..) .. Of course hard drive space has come down in pricing since then.

 

[gtomas]

Bit rate reduction, often called "compression" when referring to lossy codecs, is different from audio compression.

And so many people and engineers seem to have trouble grasping that difference!

Compression, as used in the audio processor, is the compression of the dynamic range - making softer sounds louder and (to lesser extent) making loud sounds softer. This compression is based on audio level (or loudness, if you want).

Compression, as used in the audio coding, is firstly a technically incorrect term. What should be used is reduction, because what we to do is data reduction. We throw away bits and reduce the amount of digital data, that should in the end subjectively contain the same perceived audio information. Whether the result is subjectively the same as the original audio (or exhibits distortions and what we call artifacts) depends on how much data we throw away, the quality of the codec algorithm used and the person who listens. Because we reduce the amount of data of digital audio, and in case of a finite audio clip in a file, this results in a smaller audio file, people colloquially use the term compression. "I've compressed the file, so it's smaller." But you have actually reduced the digital data (and thrown the "excess" away irreversibly) which results in a smaller file.

But these two compressions are completely different and serve different purposes!


Regards,
Goran Tomas

 

[gtomas]

If the station has analog audio processing and analog STL, high bit rate MP3 files would probably be fine.

Analog or digital processing - it doesn't matter, it's the same! Analog or uncompressed digital STL - again, it doesn't matter, it's the same. Compressed digital STL, as you mention, is different.

Analog processor are no less forgiving to coded audio, than digital audio processors are. They will both do multi-band leveling and gain control, which dynamically equalizes the program material (sometimes drastically so). The perceptual coding relies on removing acoustical information that is masked and we shouldn't hear it. But when you change the frequency balance of the perceptually coded audio, what should be masked may not be masked any more and what shouldn't be masked, may get masked. And so, the audio processor reveals artifacts that are unnoticeable in the original audio file!

Next, audio processor will employ pre-emphasis to the high frequencies. Ask yourself, when you listen to the perceptually coded audio such as bad MP3, where are the most artifacts noticeable? Are they in the low end (bass) or in the mid and high band? Where's the swishing, swooshing and the gritty sound? Now if we apply up to 17 dB of boost there, I'm wondering whether this will make these artifacts (that may have been below perceptually objectionable threshold) more noticeable or... But hey, we don't stop there! We further limit it, clip it and add distortion.

Every audio processor will reveal much more artifacts and distortion in the perceptually coded audio than it's noticeable in the original audio! That's why an MP3 may not sound bad when you listen to it on it's own (because things are masked) but what comes out on-air is no longer fine. I like to say that the quality of perceptual coded audio is fragile. On it's own, it may sound OK. Do any processing on it and it will reveal it's ugly head! Unlike uncompressed, linear audio that is infinitely more robust to any audio processing.

If you ask me, MP3 audio files, even high quality, don't cut it. If you care about your sound and quality of your end product (broadcast), linear PCM files are the only way to go.

Now we come up to the problem of cascading codecs, that you mention with the STLs. There are uncompressed digital STLs that are transparent in the audio quality sense. And there are compressed ones, that are not, because they use perceptual coding to reduce the data rate to be able to push more channels, or to push audio through a smaller data pipe. Everything above, applies here. The difference being you don't use perceptual coding on your source file, but you use it in transmission. But the same thing happens in the audio processor. There are some additional caveats depending on where the audio processor is in the chain, but we'll skip that here.

However, if you use perceptually coded source audio such as MP3s AND you use compressed digital STL, you are in double trouble because you are applying perceptual coding twice. Depending on the bitrates, this quickly degrades audio, but in case is a bad thing to do. Oh and, you will also process that audio, so add all the above in addition!

Btw, the same thing happens when you are broadcasting digital/HD radio and streaming. Here, the perceptual coding is applied at the end of the transmission chain - in the HD radio exciter or stream encoder. If you have your source files in MP3 and broadcast HD radio or stream, you are perceptually coding (and I will say, degrading) audio twice. If you use MP3s, have a compressed digital STL and are broadcasting HD radio, then it's a triple encoding pass. If your source MP3 did not come directly off the CD, but may have been transcoded in the past (that you don't know off), edited from MP3 and encoded back to MP3 (which happens so often) then you are in multiple passes of encoding and will sound really bad.

A solution to all of this - very simple! Use uncompressed, linear PCM audio that comes for a reliable source! And avoid all the problems above...


Regards,
Goran Tomas

 

[TomT]

Goran:

MP3 will remain in use, by necessity, for commercials, news, weather. As you note, the artifacts are most notable on high frequency material. Not as much a problem with voice.

The days of having spots sent in on CD from the agency, or running around town collecting local copy from other stations, are long gone. Everything is a download or e-mail. Our weather forecasts are downloaded from AccuWeather in Pennsylvania; we ship school closing from our live station to the satellite fed station by internet (Go-to-my PC).

More of a problem are those syndicated shows by download--e.g. A.T. 40 70's/80's replay. Also a problem is the store and play function of the XDS receivers--we have problems with Tom Kent doing this.

I'm a fan of good digital processing compared to the 8100/XT. Much more flexible. However, I still prefer a good analog console over digital--unless studio constrains absolutely demand digital systems. A stand-alone digital console is a waste of money. A well engineered analog STL is inherently more reliable than digital, (we have a 25 Mile path to one station--using a 606C). Less is more when it comes to digital conversions.